Brain Struct Funct DOI 10.1007/s00429-012-0418-1

ORIGINAL ARTICLE

Volumetric and surface-based 3D MRI analyses of fetal isolated mild ventriculomegaly Brain morphometry in ventriculomegaly Julia A. Scott • Piotr A. Habas • Vidya Rajagopalan Kio Kim • A. James Barkovich • Orit A. Glenn • Colin Studholme

•

Received: 23 November 2011 / Accepted: 10 April 2012 Ó Springer-Verlag 2012

Abstract Diagnosis of fetal isolated mild ventriculomegaly (IMVM) is the most common brain abnormality on prenatal ultrasound. We have set to identify potential alterations in brain development specific to IMVM in tissue volume and cortical and ventricular local surface curvature derived from in utero magnetic resonance imaging (MRI). Multislice 2D T2-weighted MRI were acquired from 32 fetuses (16 IMVM, 16 controls) between 22 and 25.5 gestational weeks. The images were motion-corrected and reconstructed into 3D volumes for volumetric and curvature analyses. The brain images were automatically segmented into cortical plate, cerebral mantle, deep gray nuclei, and ventricles. Volumes were compared between IMVM and control subjects. Surfaces were extracted from the segmentations for local mean surface curvature measurement on the inner cortical plate and the ventricles. Linear models were estimated for age-related and ventricular volume-associated changes in local curvature in both the inner cortical plate and ventricles. While ventricular volume was enlarged in IMVM, all other tissue volumes were not different from the control group. Ventricles increased in curvature with age along the atrium and anterior body. Increasing ventricular volume was associated with reduced curvature over most of the ventricular surface. The cortical plate changed in curvature with age at

J. A. Scott (&)
P. A. Habas
V. Rajagopalan
K. Kim
C. Studholme Biomedical Image Computing Group, Departments of Pediatrics, Bioengineering, and Radiology, University of Washington, Seattle, WA 98195, USA e-mail: [email protected] A. J. Barkovich
O. A. Glenn Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA 94143, USA

multiple sites of primary sulcal formation. Reduced cortical folding was detected near the parieto-occipital sulcus in IMVM subjects. While tissue volume appears to be preserved in brains with IMVM, cortical folding may be affected in regions where ventricles are dilated. Keywords Fetal MRI
Brain development
Ventriculomegaly
Cortical folding
Parieto-occipital sulcus

Introduction Ventriculomegaly, which occurs in 0.15–0.7 % of pregnancies, is often associated with other central nervous system (CNS) abnormalities, extra-CNS anomalies, chromosomal abnormalities, and developmental delay (Wax et al. 2003). As the degree of ventriculomegaly becomes more severe, the likelihood of these comorbidities increases (Falip et al. 2007; Weichert et al. 2010). While normal ventricular atrial diameter lies between 6 and 9 mm, diameters between 10 and 15 mm are considered mild ventriculomegaly. In conjunction with genetic screening to assess chormosomal abnormalities, fetal magnetic resonance imaging (MRI) has become a helpful tool in detecting CNS abnormalities to provide a more comprehensive diagnosis when ventriculomegaly is suspected (Gaglioti et al. 2009; Senapati et al. 2010; Dhouib et al. 2011). Cases with mild ventriculomegaly between 10 and 12 mm on both ultrasound and fetal MRI without any other anomalies are generally associated with good outcomes, though some will have abnormal outcomes (Wax et al. 2003; Signorelli et al. 2004; Ouahba et al. 2006; Falip et al. 2007; Weichert et al. 2010). Clearly distinguishing between those with isolated mild ventriculomegaly (IMVM)

123

Brain Struct Funct

that will have normal outcome with those that will have abnormal outcomes is therefore critical in counseling pregnancies (Guibaud 2009a). Routine assessment of ventriculomegaly relies on ventricular atrial diameter measured on ultrasonography and fetal MRI (Glenn and Barkovich 2006; Gaglioti et al. 2009; Guibaud 2009b). Ventriculomegaly has also been assessed by ventricular volume estimated from 2D in utero MRI (Grossman et al. 2006; Kazan-Tannus et al. 2007). In a small sample of non-isolated ventriculomegaly, Grossman and colleagues did not find any differences in supratentorial volume in this patient group. Likewise, a study by Kazan-Tannus used scans from 50 patients with various forms of ventriculomegaly to estimate the growth trajectory of the fetal brain from 17 to 37 gestational weeks (GW). The supratentorial volumes reported in these studies were similar to studies of control fetal populations (Scott et al. 2011; Gholipour et al. 2011). These few studies suggest that brain volume may not be significantly altered in isolated ventriculomegaly. In addition to variation in growth rate of the brain, ventriculomegaly may also be associated with disturbances Fig. 1 Qualitative example of differences in cortical plate appearance in typically developing (a and c) and isolated mild ventriculomegaly (VM) (b and d). a and b At the same gestational age, here 23 gestational weeks (GW) estimated by last menstrual period, the parieto-occipital sulcus (arrowhead) is clearly less pronounced in the VM brain. c and d In the presence of dilated temporal horns, the position of the hippocampus is deviated medially, which affects the shape of the medial temporal cortical surface (*)

123

in cortical folding during the second half of gestation (Fig. 1; Habas et al. 2011). In the absence of a cortical malformation, apparent delays may be related to mechanical interference with the infolding of the cortical plate caused locally by the dilation of the ventricles, rather than abnormal corticogenesis. Investigation of the morphological changes of IMVM brains during the period of primary sulcogenesis (20–28 GW) may help illuminate how the local folding patterns of the cortical plate are affected by ventirculomegaly alone. To further address the question of whether we could detect any alterations in brain development associated with ventriculomegaly, we have compared a group of IMVM (atrial diameter between 10 and 12 mm) with normally developing fetuses between 22 and 25.5 GW. We automatically segmented the supratentorial brain using a validated, atlas-based approach on 3D reconstructed, motion-corrected T2-weighted MRIs acquired in utero (Habas et al. 2010a, b). From these segmentations, we measured tissue and ventricular volumes and also extracted the surface meshes for local curvature analysis of the ventricles and inner cortical plate. These quantitative

Brain Struct Funct

measures will address the following questions: (1) do supratentorial tissue volumes differ in IMVM, (2) where are the lateral ventricles commonly dilated in IMVM, and (3) are alterations in cortical plate surface curvature associated with enlargement of the ventricles?

Materials and methods Subjects and fetal image acquisition Data used in the present paper were acquired from an imaging study that was approved by our local Institutional Review Board and complies with NIH human subject guidelines. All enrolled mothers provided informed consent for use of fetal scans for research purposes. For control subjects, mothers were referred for fetal MRI either to better assess questionable findings on prenatal ultrasound (N = 3), or because of a prior abnormal pregnancy (N = 6) or because the patient volunteered for the study (N = 7). The IMVM group was comprised of fetuses with sonographically diagnosed IMVM (N = 16), with ventricular atria between 10 and 12 mm, who were referred for clinical fetal MRI. Mothers enrolled in the studies ranged in age from 19 to 41 years (mean, 34 years). Of the study scans that met image acquisition criteria for image analysis (TR = 3,000 - 9,000 ms; slice thickness & 3 mm; sufficient number of stacks; fetal motion less than 15 mm translation and 30° of rotation between slices), we selected 32 clinical scans, 16 controls (8 males, 8 females) and 16 IMVM (7 males, 6 females, 3 unknown) at gestational ages ranging from 22 to 25.5 GW, estimated by last menstrual period. All of the fetal MRI used in the control group were read as normal by a fetal neuroradiologist (O.A.G.). Subjects in the IMVM group had no other findings on fetal MRI. All control subjects had normal postnatal neurodevelopment, as determined by either phone questionnaire or by formal neurologic evaluation at 1 year of age. Neither the mother nor fetus was sedated and the mother was free-breathing for the scan duration of 20 to 45 min. Clinical MR imaging was performed on a 1.5T scanner (GE Healthcare, Milwaukee, WI) using an eight-channel torso phased-array coil. Multiple stacks of single-shot fast spin-echo (SSFSE) T2w slice images (in-plane pixel size & 0.5 9 0.5 mm; slice thickness & 3 mm, no gap) were planned in the approximately axial, sagittal, and coronal planes with respect to the fetal brain. The underlying spatial resolution of the original 2D multi-slice data was between 0.8 and 1.0 mm, prior to clinically requested re-gridding to between 0.4 and 0.5 mm in-plane pixel size for radiological inspection on clinical PACS. All slice images were acquired in an interleaved manner to reduce saturation of spins in adjacent slices. The MR sequence

parameters (TR = 3,000 - 9,000 ms; TE = 91 ms) were originally designed for clinical scans and were not adjusted for tissue segmentation in this study. From these clinical acquisitions, one to four stacks were selected in each plane for 3D image reconstruction. To account and correct for spontaneous fetal movement during scanning, all image slices in the slice stacks of a subject were registered using the slice intersection motion correction (SIMC) technique (Kim et al. 2010) and reconstructed into 3D volumes with isotropic voxel dimensions of 0.5 mm. Volumetric analysis The reconstructed MR volumes were automatically segmented, using an atlas-based approach (Habas et al. 2010a, b), into the following supratentorial structures: cortical plate (CP); cerebral mantle (CM) which includes subplate, intermediate zone, and germinal matrix; deep gray (DG) which includes basal ganglia and thalamus; and ventricles (VENT) including lateral and third ventricles. A computational atlas of MRI intensity, tissue probability, and shape of the fetal brain was previously generated from manual segmentations of 30 normal fetal brain anatomies between 20 and 30 GW (Habas et al. 2010a; Scott et al. 2011). From this continuous model of fetal brain growth, a synthetic age-specific MRI intensity template and an agespecific tissue probability map were generated for the gestational age of each subject to be analyzed. After the subject MRI was aligned to the age-matched MRI template using non-linear registration, the age-matched tissue probability map was used as a source of spatial priors for automatic atlas-based segmentation of developing brain tissues (Habas et al. 2010b). This automatic segmentation approach was previously validated for young fetuses with normal brain development (Habas et al. 2010a, b; Scott et al. 2011). For additional validation for IMVM subjects, manual segmentations of five IMVM scans were compared by percent overlap to the corresponding automated segmentations. The average Dice coefficients for all tissue zones were above 0.8 (CP = 0.84, CM = 0.91, DG = 0.81, and VENT = 0.88), indicating very good agreement between manual and automatic tissue delineations. Examples of the automated segmentations in control and IMVM brains are shown in Fig. 2a–b. Linear estimations of volume change by group (IMVM and CON) with age were calculated for each segmentation label and supratentorial volume (sum of CP, CM, and DG). The interaction of age and group was not significant for all structures, so a common slope was used in each model and intercept differences by group were tested. Hemispheric asymmetry in each tissue and ventricular volumes were tested within group by a t test.

123

Brain Struct Funct Fig. 2 Automated tissue segmentation and atrial diameter. a Automated and b manual segmentation of supratentorial tissues in a VM case. The supratentorial volume was delineated into cortical plate (CP, red), cerebral mantle (CM, blue), deep gray nuclei (DG, yellow); ventricles (VENT, purple). Ventricular atrial diameter was measured in the axial plane about the superior aspect of the thalamus in both (c) control and (d) VM subjects

Right and left ventricular atrial diameters were also manually measured on the 3D volumes for correlative analysis with the right and left ventricular volumes, respectively. The diameter was measured on an axial slice of the ventricular atrium at the level of the glomus of the choroid plexus, near the superior aspect of the thalamus (Fig. 2c–d). Local curvature analysis Local curvature analysis methodology executed in this study is based on the methods used in a study of a control fetal population (Habas et al. 2012). First, tissue maps extracted from automatic segmentations were smoothed using 2 mm full width at half-maximum (FWHM) Gaussian kernel and linearly registered to a reference anatomy. Then, the tissue map of each subject was tessellated into two triangular meshes using a topology-preserving marching cubes algorithm (Lopes and Brodlie 2003) to

123

reconstruct the inner cortical plate surface and the ventricular surface. Local geometry of each surface was quantified using mean surface curvature H calculated at each mesh vertex (Do Carmo 1976). Mean curvature provides a physically interpretable measure of surface geometry—positive values of H express the convexity of a region, whereas negative values of H indicate the level of local concavity. For the purpose of modeling of temporal shape changes and curvature differences caused by ventricular enlargement, measurements of local surface curvature were mapped from individual subject surfaces onto populationaverage surfaces obtained by group-wise volumetric registration of the tissue maps (Studholme and Cardenas 2004; Rajagopalan et al. 2011). We tested the effect of ventriculomegaly in two separate models. First, diagnosis was treated as a group factor (aIMVM). For each vertex x of the reference surface, age- and group-related changes in local surface curvature were represented using a linear model

Brain Struct Funct

H(x,t) = bage(x)t ? bgroup(x)aIMVM ? b0(x) with coefficients found through least squares fitting to measurements of curvature mapped from all study subjects. In this model, t is gestational age in weeks and aIMVM = 0 is the control group and aIMVM = 1 is the patient group. In the second model, ventricular volume was entered as a covariate (vVENT) because this continuous variable did not overlap between control and IMVM groups and represented degree of ventricular enlargement independent of diameter-based diagnosis. Age- and ventricular volume-related changes were represented using a linear model H(x,t) = bage(x)t ? bvol(x)vVENT ? b0(x). Here, t is gestational age in weeks and vVENT is total ventricular volume. Statistical significance with multiple comparisons correction of the terms bage(x), bgroup(x) and bvol(x) were assessed using permutation testing in each model (Nichols and Holmes 2002).

Results Volumetric analysis All automated tissue segmentations were visually inspected (J.A.S.) for anatomical accuracy of tissue labels. An example of the automated segmentation in an IMVM subject is shown in Fig. 3. Total ventricular volume was significantly greater at all ages in the IMVM group compared with controls (p \ 0.001) (Fig. 4). On average, ventricular volume was enlarged by 3.36 cm3 (111 %) in the IMVM group. In contrast, total supratentorial volume did not significantly

Fig. 3 Automated segmenation of an IMVM subject. Visual inspection insured that the automated segmenations performed well in all areas of the brain. For example, the cortical plate of the medial wall is accurately labeled along the medial temporal cortical plate (b) and the

differ by group (Fig. 4). Group comparison of cortical plate, cerebral mantle, and deep gray volumes showed no significant differences (Fig. 4). Though range in atrial width was narrow in the IMVM group, ventricular volume had greater variance compared with controls (IMVM = 3.17 cm3, CON = 0.61 cm3). In either group, ventricular volume did not significantly change with age (Fig. 4). Total supratentorial volume increased at the rate of 13.35 cm3/wk for the total population (p \ 0.001). The ratio of each tissue volume to supratentorial volume did not change from 22 to 25.5 GW. However, the ratio of ventricular volume to supratentorial volume did decrease (p \ 0.001), as would be expected when ventricular volume does not change and suprtentorial volume increases significantly. Also, this ratio was significantly greater in the IMVM group compared with controls (p \ 0.001). Ventricular volume had greater hemispheric asymmetry (7 %) than cerebral tissue volumes (\2 %). Two control brains and two IMVM brains had highly rightward asymmetric ventricles with an absolute asymmetry coefficient greater than one standard deviation above the mean. In addition, one IMVM brain had leftward asymmetry. The IMVM subjects with highly asymmetric ventricular volumes were also diagnosed as unilateral IMVM by atrial diameter measurement on 2D MRI. Ventricular atrial diameter manually measured on 3D MRI ranged from 4 to 12 mm for the whole population. The IMVM group diameters fell between 10 and 12 mm in one or both hemispheres (average, 10.4 mm) and the control group diameters were less than 10 mm in both hemispheres (average, 6.6 mm). The correlation between

parieto-occipital sulcal region (a, c). Also the corpus callosum is included within cerebral mantle label over the entire length of the structure (c)

123

Brain Struct Funct Fig. 4 Volumetric growth trajectories for controls (open circles) and VM (closed circles) groups. Ventricle volume was greater in the VM group compared to controls over the entire age range (22–25.5 GW). Volume on average did not differ between controls and VM for total tissue volume (ST) or any individual tissues (CP, CM, or DG). For all tissues and ventricles, the groups did not differ in the rate of volume increase either

hemispheric atrial diameter and ventricular volumes were significant (Fig. 5) (right, R2 = 0.44, p \ 0.001; left, R2 = 0.57, p \ 0.001). Although volume reflects the total ventricular size and diameter only directly measures ventricle size in the atrium, volume still shows a strong relationship with diameter and so we have used ventricular volume as a covariate for surface curvature analysis of the ventricles and cortical plate.

123

Ventricular curvature Change in ventricular surface curvature was linearly modeled for the effect of age and ventricular volume on the entire population (controls and IMVM). From 22 and 25.5 GW, the ventricles increased in convexity bilaterally along the anterior and posterior aspects of the atrium and the inferior aspect of the anterior body (Fig. 6a). The remainder

Brain Struct Funct Fig. 5 a Left and b right ventricular volumes and atrial diameters in control (open circles) and VM (closed circles) groups. In both hemispheres, volume and diameter were significantly correlated (p \ 0.001)

Fig. 6 T maps on the average ventricular surface representing significant (a) age-related changes in curvature and (b) ventricular volume associated curvature. a Agerelated increases in convexity (sharper angles) are localized to the anterior and posterior aspects of the atrium and the inferior aspect of the anterior body of the lateral ventricles. b There is a negative association between ventricular volume and ventricular curvature over most of the surface

123

Brain Struct Funct

of the ventricular surface did not change in local surface curvature with age. Increasing ventricular volume was associated with bilateral reduction in convexity (i.e., increased dilation) over much of the lateral ventricular surface (Fig. 6b). In particular, the lateral and medial aspects of the atrium and the temporal horns decreased in convexity. Contrast between the IMVM and control groups yielded a very similar pattern of curvature reductions due to the effect of IMVM diagnosis (data not shown). Ventricular dilation in IMVM was not limited to the ventricular atrium, where ventriculomegaly diagnosis was made. The atrium was also an area that changed in curvature with age, and therefore appearance, during the diagnostic period for IMVM. Though the third ventricle was included in our ventricle definition, it did not show any significant changes in local surface curvature with age or increasing ventricular volume. Cortical plate curvature Age-related and ventricular enlargement-associated changes in cortical plate local surface curvature were detected over the whole population. Over this period of gestational development, progression in cortical folding was evident (Fig. 7a). At the site of the central sulcus in both hemispheres, the cortical plate surface increased in concavity. Anterior and posterior to the superior extent of the central sulcus were increases in convexity corresponding to the precentral and the postcentral gyri, respectively. The superior frontal gyri and sulci were also forming at this time with adjacent increased concavity and increased convexity in the frontal lobe. Progress in the closure of the Sylvian fissure was also detected. This was shown by the increasing convexity of the opercular cortices and decreasing convexity of the posterior insular cortex. Last, the parieto-occipital sulcus increased in concavity. Compared with the left hemisphere, the right hemisphere had a significantly greater area of the cortical surface that changed in curvature, which would be expected from observed curvature asymmetries in the control population (Habas et al. 2012). In association with greater ventricular volume or IMVM diagnosis, reduced cortical folding (decreased concavity at sulcal sites) was detected bilaterally along the anterior aspect of the inferior parieto-occipital sulcus near the junction with calcarine sulcus (Fig. 7b). The two models (diagnosis and ventricular volume) had overlapping areas of significantly reduced curvature (data not shown). The inferior parietooccipital sulcus lies adjacent to a region of common ventricular dilation in IMVM (Fig. 6b) and is visibly more shallow (Fig. 1). No other areas of the cortical plate were significantly affected by ventricular volume bilaterally.

123

Discussion To date, this is the first study of tissue-specific growth trajectories and cortical surface curvature in fetal ventriculomegaly. Our methodology has shown that we are sensitive to differences in brain development between fetuses with IMVM and normal fetuses. The comparison of age-matched IMVM and control fetal brains has shown remarkable similarities and focal differences late in the second trimester of gestation. Ventricular enlargement (by volume and dilation) and alteration in the cortical folding along the anterior aspect of the parieto-occipital sulcus distinguished fetuses with IMVM from normal fetuses. We have shown that the dilation of the ventricles extends beyond the atrium even in the mildest form of ventriculomegaly. Also bilateral IMVM ventricular volumes did not overlap with age-matched control subjects, demonstrating the strength of ventricular volume as an accurate assessment of ventriculomegaly. While IMVM brain morphology is distinct, supratentorial and tissue volumes are normal in midgestation. Our findings agree with past studies that have reported similar global brain volumes between ventriculomegaly and control groups (Grossman et al. 2006; KazanTannus et al. 2007). Our study adds to past imaging studies by showing that the cortical plate, the cerebral mantle, and the deep gray nuclei each have normal volumes between 22 and 25.5 GW. Although most fetuses with IMVM have a normal neurodevelopmental outcome, a minority can have developmental disabilities (Signorelli et al. 2004; Ouahba et al. 2006; Falip et al. 2007; Melchiorre et al. 2009; Beeghly et al. 2010). The relationships between fetal diagnosis and postnatal brain structural maturation is not explored in studies of fetal IMVM, which leaves open the question of what underlies the developmental abnormalities that are observed in some patients. One postnatal neuroanatomical study by Gilmore and colleagues has shown that neonates with IMVM have larger intracranial volumes and greater gray to white matter ratios compared with controls (Gilmore et al. 2008). In our study of the fetal brain late in the second trimester, cerebral volume and cortical plate to cerebral mantle volume ratios were similar between IMVM and control groups. This suggests that there may be differences that still need to be investigated in the developmental trajectory in the third trimester as the cortical sheet growth accelerates and cortical connections are being established. While some severe forms of ventriculomegaly have cortical malformations that are associated with abnormal postnatal development (Glenn and Barkovich 2006; Gaglioti et al. 2009), subtle alterations in IMVM cortical folding patterns may also be associated with poorer outcomes. Li and colleagues have associated the visualization of the

Brain Struct Funct Fig. 7 T maps on the average cortical plate surface representing significant (a) agerelated changes in curvature and (b) ventricular volumeassociated curvature. Increases in convexity are shown in cool colors and decreases are warm colors. a Age-related changes in curvature were detected bilaterally in the perislyvian region (SF), about the central sulcus (CS), the posterior part of the superior frontal sulcus (SFS), and the parieto-occipital sulcus (POS). b Greater ventricular volume was associated with reduced concavity (increased convexity) at the POS near the junction with the caclcarine sulcus (CaS)

parieto-occipital sulcus, in addition to the cingulate and superior temporal sulci, on routine clinical MR images with the increased likelihood of normal postnatal outcome in fetuses with isolated ventriculomegaly of varying severity (Li et al. 2011). Future study of a larger IMVM population with the 3D morphometric techniques used here could show associations between outcome and local cortical folding patterns, independent of qualitative assessments. In our study, alterations in cortical folding associated with dilated lateral ventricles were only significant in the

medial occipito-parietal region of the cortical plate. The same region (POS/CaS) was also bilaterally reduced in curvature when the IMVM group was strictly contrasted with controls (Habas et al. 2011). Though a single focal area was commonly reduced in curvature in our narrow classification of IMVM, other sulci could potentially be associated with ventricular dilation. Expansion of the study of sulcogenesis to IMVM at later gestational ages may reveal other cortical areas that are consistently affected by dilation of the lateral ventricles.

123

Brain Struct Funct

Though local curvature analysis of cortical folding may demonstrate where variations occur due to ventricular enlargement, it does not explain the mechanism causing the physical phenomenon. This may be a more general phenomenon in which dilated ventricles result in shifts of tissue outward rather than a sign of delayed or abnormal corticogenesis (Kanekar and Gent 2011). Uncovering this extended pattern may require larger sample populations. The mechanisms behind cortical folding disturbances may be of particular interest in more severe and non-isolated forms of ventriculomegaly, which are associated with poorer neurodevelopmental outcomes (Signorelli et al. 2004; Weichert et al. 2010), as well as in IMVM.

Conclusions Previous in utero studies of brain development in ventriculomegaly have been limited to 2D images from ultrasonography (Wax et al. 2003) or 2D MRI (Grossman et al. 2006; Kazan-Tannus et al. 2007). In this study, we have taken a more comprehensive evaluation of brain morphometry in fetal IMVM with 3D reconstructed MRI. We have shown that while a great area of the lateral ventricles is enlarged, brain volume, specifically that of the cortical plate, is normal. Also the shape of the cortical sheet is only slightly altered along the medial wall in an area adjacent to the ventricular atrium, the region of most common ventricular dilation in IMVM. Follow-up fetal and neonatal imaging should be carried out to determine whether brain growth maintains a normal course and how cortical folding patterns emerge. Acknowledgements This research was funded by the National Institutes of Health through the National Institute of Neurological Disorders and Stroke (R01 NS 061957 and R01 NS 055064); National Center for Research Resources to UCSF-CTSI (UL1 RR024131); and award to O.A.G. (K23 NS52506-03). Conflict of interest of interest.

The authors declare that they have no conflict

References Beeghly M, Ware J, Soul J, du Plessis A, Khwaja O, Senapati GM, Robson CD, Robertson RL, Poussaint TY, Barnewolt CE, Feldman HA, Estroff JA, Levine D (2010) Neurodevelopmental outcome of fetuses referred for ventriculomegaly. Ultrasound Obstet Gynecol 35(4):405–416 Dhouib A, Blondiaux E, Moutard ML, Billettede Villemeur T, Chalard F, Jouannic JM, Ducoule Pointe H, Garel C (2011) Correlation between pre- and postnatal cerebral magnetic resonance imaging. Ultrasound Obstet Gynecol 38(2):170–178 Do Carmo MP (1976) Differential geometry of curves and surfaces. Prentice Hall, Englewood Cliffs

123

Falip C, Blanc N, Maes E, Zaccaria I, Oury JF, Sebag G, Garel C (2007) Postnatal clinical and imaging follow-up of infants with prenatal isolated mild ventriculomegaly: a series of 101 cases. Pediatr Radiol 37(10):981–989 Gaglioti P, Oberto M, Todros T (2009) The significance of fetal ventriculomegaly: etiology, short- and long-term outcomes. Prenat Diagn 29(4):381–388 Gholipour A, Estroff JA, Barnewolt CE, Connolly SA, Warfield SK (2011) Fetal brain volumetry through MRI volumetric reconstruction and segmentation. Int J Comput Assist Radiol Surg 6(3):329–339 Gilmore JH, Smith LC, Wolfe HM, Hertzberg BS, Smith JK, Chescheir NC, Evans DD, Kang C, Hamer RM, Lin W, Gerig G (2008) Prenatal mild ventriculomegaly predicts abnormal development of the neonatal brain. Biol Psychiatry 64(12): 1069–1076 Glenn OA, Barkovich AJ (2006) Magnetic resonance imaging of the fetal brain and spine: an increasingly important tool in prenatal diagnosis, part 1. AJNR Am J Neuroradiol 27(8): 1604–1611 Grossman R, Hoffman C, Mardor Y, Biegon A (2006) Quantitative MRI measurements of human fetal brain development in utero. Neuroimage 33(2):463–470 Guibaud L (2009) Contribution of fetal cerebral MRI for diagnosis of structural anomalies. Prenat Diagn 29(4):420–433 Guibaud L (2009) Fetal cerebral ventricular measurement and ventriculomegaly: time for procedure standardization. Ultrasound Obstet Gynecol 34(2):127–130 Habas PA, Kim K, Corbett-Detig JM, Rousseau F, Glenn OA, Barkovich AJ, Studholme C (2010) A spatiotemporal atlas of MR intensity, tissue probability and shape of the fetal brain with application to segmentation. Neuroimage 53(2):460–470 Habas PA, Kim K, Rousseau F, Glenn OA, Barkovich AJ, Studholme C (2010) Atlas-based segmentation of developing tissues in the human brain with quantitative validation in young fetuses. Hum Brain Mapp 31(9):1348–1358 Habas PA, Rajagopalan V, Scott JA, Kim K, Roosta A, Rousseau F, Barkovich AJ, Glenn OA, Studholme C (2011) Detection and mapping of delays in early cortical folding derived from in utero MRI. In: Medical Imaging 2011: Image Processing, Proceedings of SPIE, vol 7962, 79,624D Habas PA, Scott JA, Roosta A, Rajagopalan V, Kim K, Rousseau F, Barkovich AJ, Glenn OA, Studholme C (2012) Early folding patterns and asymmetries of the normal human brain detected from in utero MRI. Cereb Cortex 22(1):13–25 Kanekar S, Gent M (2011) Malformations of cortical development. Semin Ultrasound CT MR 32(3):211–227 Kazan-Tannus JF, Dialani V, Kataoka ML, Chiang G, Feldman HA, Brown JS, Levine D (2007) MR volumetry of brain and CSF in fetuses referred for ventriculomegaly. AJR Am J Roentgenol 189(1):145–151 Kim K, Habas PA, Rousseau F, Glenn OA, Barkovich AJ, Studholme C (2010) Intersection based motion correction of multislice MRI for 3-D in utero fetal brain image formation. IEEE Trans Med Imaging 29(1):146–158 Li Y, Estroff JA, Mehta TS, Robertson RL, Robson CD, Poussaint TY, Feldman HA, Ware J, Levine D (2011) Ultrasound and MRI of fetuses with ventriculomegaly: can cortical development be used to predict postnatal outcome? AJR Am J Roentgenol 196(6): 1457–1467 Lopes A, Brodlie K (2003) The robustness and accuracy of the marching cubes algorithm for isosurfacing. IEEE Transact Visualiz Comput Graph 9(1):16–29 Melchiorre K, Bhide A, Gika AD, Pilu G, Papageorghiou AT (2009) Counseling in isolated mild fetal ventriculomegaly. Ultrasound Obstet Gynecol 34(2):212–224

Brain Struct Funct Nichols TE, Holmes AP (2002) Nonparametric permutation tests for functional neuroimaging: a primer with examples. Hum Brain Mapp 15(1):1–25 Ouahba J, Luton D, Vuillard E, Garel C, Gressens P, Blanc N, Elmaleh M, Evrard P, Oury JF (2006) Prenatal isolated mild ventriculomegaly: outcome in 167 cases. BJOG 113(9):1072– 1079 Rajagopalan V, Scott J, Habas PA, Kim K, Corbett-Detig J, Rousseau F, Barkovich AJ, Glenn OA, Studholme C (2011) Local tissue growth patterns underlying normal fetal human brain gyrification quantified in utero. J Neurosci 31(8):2878–2287 Scott JA, Habas PA, Kim K, Rajagopalan V, Hamzelou KS, CorbettDetig JM, Barkovich AJ, Glenn OA, Studholme C (2011) Growth trajectories of the human fetal brain tissues estimated from 3D reconstructed in utero MRI. Int J Dev Neurosci 29(5):529–536 Senapati GM, Levine D, Smith C, Estroff JA, Barnewolt CE, Robertson RL, Poussaint TY, Mehta TS, Werdich XQ, Pier D,

Feldman HA, Robson CD (2010) Frequency and cause of disagreements in imaging diagnosis in children with ventriculomegaly diagnosed prenatally. Ultrasound Obstet Gynecol 36(5): 582–595 Signorelli M, Tiberti A, Valseriati D, Molin E, Cerri V, Groli C, Bianchi UA (2004) Width of the fetal lateral ventricular atrium between 10 and 12 mm: a simple variation of the norm? Ultrasound Obstet Gynecol 23(1):14–18 Studholme C, Cardenas V (2004) A template free approach to volumetric spatial normalization of brain anatomy. Pattern Recogn Lett 25(10):1191–1202 Wax JR, Bookman L, Cartin A, Pinette MG, Blackstone J (2003) Mild fetal cerebral ventriculomegaly: diagnosis, clinical associations, and outcomes. Obstet Gynecol Surv 58(6):407–414 Weichert J, Hartge D, Krapp M, Germer U, Gembruch U, AxtFliedner R (2010) Prevalence, characteristics and perinatal outcome of fetal ventriculomegaly in 29,000 pregnancies followed at a single institution. Fetal Diagn Ther 27(3):142–148

123